pgRouting Manual (2.2)

Performance Tips

«  Dictionary of columns & Custom Query   ::   Contents   ::   User’s Recipes List  »

Performance Tips

For the Routing functions:

Note

To get faster results bound your queries to the area of interest of routing to have, for example, no more than one million rows.

For the topology functions:

When “you know” that you are going to remove a set of edges from the edges table, and without those edges you are going to use a routing function you can do the following:

Analize the new topology based on the actual topology:

pgr_analyzegraph('edge_table',rows_where:='id < 17');

Or create a new topology if the change is permanent:

pgr_createTopology('edge_table',rows_where:='id < 17');
pgr_analyzegraph('edge_table',rows_where:='id < 17');

Use an SQL that “removes” the edges in the routing function

SELECT id, source, target from edge_table WHERE id < 17

When “you know” that the route will not go out of a particular area, to speed up the process you can use a more complex SQL query like

SELECT id, source, target from edge_table WHERE
        id < 17 and
        the_geom  && (select st_buffer(the_geom,1) as myarea FROM  edge_table where id=5)

Note that the same condition id < 17 is used in all cases.

«  Dictionary of columns & Custom Query   ::   Contents   ::   User’s Recipes List  »